[jira] Updated: (DIRMINA-519) BufferingFilter

["Edouard De Oliveira (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/DIRMINA-519?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Edouard De Oliveira updated DIRMINA-519:
----------------------------------------

    Attachment: test.rar

This file contains a micro benchmark demonstrating the benefits of the LazyInitializedCacheMap commited to trunk.
Only tested under winXP on a 2way cpu under jdk1.6. The recent improvements on synchronized and the changes on the jmm seem to have affected the gap between HashMap and ConcurrentHashMap. 
But due to the cost of an IoBuffer allocation and the specific use of this buffering filter, on my configuration it seems accurate to use this special map.

Need other testers to confirm it.

> BufferingFilter
> ---------------
>
>                 Key: DIRMINA-519
>                 URL: https://issues.apache.org/jira/browse/DIRMINA-519
>             Project: MINA
>          Issue Type: New Feature
>          Components: Filter
>            Reporter: Trustin Lee
>            Assignee: Edouard De Oliveira
>            Priority: Minor
>             Fix For: 2.0.0-M3
>
>         Attachments: test.rar
>
>
> As JDK provides BufferedOutputStream, we could provide BufferingFilteer which does the same thing, which buffers encoded data and flushes it out when the buffer becomes full or the flush operation is explicitly requested.  This kind of filter is sometimes useful when a session is generating very small messages too frequently and consequently generates unnecessary traffic overhead.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.